如何折疊 R 中分類變量的水平

Question

我有各種分類變量，每個變量都超過 5 個級別，我想要一個可以將它們折疊成兩個級別的函數

column1<- c("bad","good","nice","fair","great","bad","bad","good","nice",
            "fair","great","bad")
column2<- c("john","ben","cook","seth","brian","deph","omar","mary",
            "frank","boss","kate","sall")

df<- data.frame(column1,column2)

因此，對於上面的數據框，在 column1 中，我想使用一個函數將所有“壞”轉換為“壞”，將其他級別轉換為“其他”。 我不知道該怎么做。 謝謝

Answer 1

使用ifelse或case_when

library(dplyr)
df <- df %>% 
   mutate(column1 = case_when(column1 != "bad" ~ "others", TRUE ~ column1))

---

此外，由於只有一個變化，我們可以做

df$column1[df$column1 != "bad"] <- "others"

Answer 2

在 base R 中執行此操作的一種簡單方法是使用索引：

c('others', 'bad')[(df$column1 == 'bad') + 1]
#> [1] "bad"    "others" "others" "others" "others" "bad"    "bad"   
#> [8] "others" "others" "others" "others" "bad"  

Answer 3

df<- data.frame(factor=as.factor(column1),column2)
levels(df$factor)<-c("bad",rep("other",4))

Answer 4

這是帶分組的dplyr解決方案：

library(dplyr)
df %>% 
  group_by(group = cumsum(column1=="bad")) %>% 
  mutate(column1 = ifelse(row_number()==1, "bad", "others")) %>% 
  ungroup() %>% 
  select(-group)

  column1 column2
   <chr>   <chr>  
 1 bad     john   
 2 others  ben    
 3 others  cook   
 4 others  seth   
 5 others  brian  
 6 bad     deph   
 7 bad     omar   
 8 others  mary   
 9 others  frank  
10 others  boss   
11 others  kate   
12 bad     sall